A new project called ContextMiner has been created by researchers at the University of North Carolina at Chapel Hill. The tool lets anyone automate the collection of links to online videos and blogs along with their extensive metadata. Although they're calling ContextMiner a YouTube archiving tool, it doesn't actually download the videos off the site...yet. Instead, it extracts the embed, and the provides that to you along with other details like the number of views and what sites are linking to the video.

The tool, a part of the university's NDIIPP VidArch project, is designed to be a framework that collects, analyzes, and presents contextual information along with the data it archives. To get started with ContextMiner, you create a scheduled, repeated collection activity called a "campaign." For each campaign, you can enter in details like description and scope, then customize how often the campaign should run (daily, weekly, monthly), among other things. If you want to collect "in-links" - the web sites on the internet linking to the video in question - that is also an option. In addition to scouring YouTube, you can configure ContextMiner to search through the web and blogs, too.

Why Use ContextMiner?

Marketers will probably be interested in how this free tool is able to track views and links, but that's not really the purpose behind ContextMiner's creation. Instead, the tool is designed more for research than anything else. For example, one of the main reasons to use ConextMiner is its ability to document the cultural phenomena of viral videos.

Often, when a video goes viral, very few people are aware of where it came from, what the story is behind it, who created it and why. As time goes by, finding the original video creator and source is even harder as the video spreads across the internet. But now, thanks to ContextMiner, the history behind a video's creation is no longer a mystery.

Take for instance, Vote Different, the mashup of Hillary Clinton with the famous Apple 1984 Super Bowl ad and one of the most popular videos on YouTube. We've probably all seen this video at one point or another, but did you ever want to know who created it and why?

With ContextMiner, that information can easily be discovered. Because of its ability to pull the inbound links to a video, we can see that the original creator of the video, a user by the name of "ParkRidge47," is the subject of one of the inbound links to the video. A blog post on TechPresident titled "Who is ParkRidge47?" gives us a great history of this particular video's creation. You could also sort through the links provided to find the very first person to link to the video, which is often the creator themselves.

ContextMiner is still under development. In the future, the developers hope to offer tools and policies for exporting the videos, blog pages, and metadata. That's probably not an empty promise - there's already an an option to "download Flash video from YouTube" on the campaign creation form, it's just disabled right now. When that feature becomes available, we think it would be fine to then call ContextMiner a YouTube archiving tool. Since "Archive" implies making a backup copy, until then we think ContextMiner should really just be considered a research tool. Still, we have to say, it's a pretty good one.